sakemin / audiosr-long-audio

Versatile Audio Super-resolution at Scale which upsamples audio files to 48khz. Longer audio input is possible with this model

  • Public
  • 2.4K runs
  • L40S
  • GitHub
  • Paper

Input

Video Player is loading.
Current Time 00:00:000
Duration 00:00:000
Loaded: 0%
Stream Type LIVE
Remaining Time 00:00:000
 
1x
*file

Audio to upsample

boolean

Truncate batches to 5.12 seconds. Essential with long audio files, due to memory issues.

Default: true

integer
(minimum: 10, maximum: 500)

Number of inference steps

Default: 50

number
(minimum: 1, maximum: 20)

Scale for classifier free guidance

Default: 3.5

integer

Random seed. Leave blank to randomize the seed

Output

Video Player is loading.
Current Time 00:00:000
Duration 00:00:000
Loaded: 0%
Stream Type LIVE
Remaining Time 00:00:000
 
1x
Generated in

This output was created using a different version of the model, sakemin/audiosr-long-audio:44b37256.

Run time and cost

This model costs approximately $0.10 to run on Replicate, or 10 runs per $1, but this varies depending on your inputs. It is also open source and you can run it on your own computer with Docker.

This model runs on Nvidia L40S GPU hardware. Predictions typically complete within 107 seconds.

Readme

AudioSR: Versatile Audio Super-resolution at Scale

Pass your audio in, AudioSR will make it high fidelity!

Work on all types of audio (e.g., music, speech, dog, raining, …) & to 48khz.

Longer audio input is also compatible with this model, by implementing audio slicing into chunks.